Scholarlink

With the intention of integrating EEG and ECG readings, this work examines multimodal deep-learning techniques for early MDD detection. Eight significant studies have shown that multimodal approaches frequently outperform single-modal ones, achieving accuracy rates. The accuracy rates mostly range from 82.68% to 99.24%. Important facts have been discovered showing that MDD patients have abnormalities in cardiovascular variability and brain network structure. Experiments using DNN, EDL, and LSTM-RNN on the Framingham Heart Study dataset reveal variations in precision, recall, and training speed but comparable accuracy (83.73–84.67%). Although encouraging, issues with model interpretability, dataset variation, and the requirement for contextual knowledge continue to exist.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Multimodal Techniques for Early Detection of Major Depressive Disorder

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Ketan Desale,
  • Vedant Gaikar,
  • Sakshi Ghardale,
  • Riya Jadhav,
  • Kartik Girase

摘要

With the intention of integrating EEG and ECG readings, this work examines multimodal deep-learning techniques for early MDD detection. Eight significant studies have shown that multimodal approaches frequently outperform single-modal ones, achieving accuracy rates. The accuracy rates mostly range from 82.68% to 99.24%. Important facts have been discovered showing that MDD patients have abnormalities in cardiovascular variability and brain network structure. Experiments using DNN, EDL, and LSTM-RNN on the Framingham Heart Study dataset reveal variations in precision, recall, and training speed but comparable accuracy (83.73–84.67%). Although encouraging, issues with model interpretability, dataset variation, and the requirement for contextual knowledge continue to exist.